Arterial stiffness and endothelial dysfunction in HIV-infected children

Background: The role of antiretroviral therapy in acceleration of atherosclerosis in HIV-infected adults is controversial, partly because of the confounding effects of the involvement of classic cardiovascular risk factors.

Objective: To study vascular function in HIV-infected children.

Design: Cross-sectional study of 49 HIV-infected children (34 receiving antiretroviral therapy and 15 never treated) and if 24 age- and sex-matched controls.

Methods: Automatic, computerized, ultrasonic procedure evaluation of geometric and mechanical properties of the common carotid artery, and of the endothelium-dependent dilation and endothelium-independent dilation.

Results: Relative systolodiastolic variations in diameter of the carotid artery in HIV-infected children were significantly lower than those in controls, but there was no significant difference in intima-media thickness. Cross-sectional compliance and distensibility were also significantly lower. Wall stiffness, assessed as the incremental elastic modulus, was larger in HIV-infected children. Endothelium-dependent dilation was lower in HIV-infected children but non-endothelium-dependent dilation was similar to that in controls. We did not find differences for any of the vascular variables between HIV-infected children receiving antiretroviral therapy and those never treated. All arterial variables were similar in children with and without dyslipidemia.

Conclusions: HIV-infected children had a vascular dysfunction in the absence of cardiovascular risk factors. In this short series, no additional detrimental effects were observed after a mean of 5 years of antiretroviral therapy.

Introduction

Since highly active antiretroviral therapy was introduced the life expectancy of HIV-infected children has increased [¹]. Consequently, the spectrum of HIV-related concerns has shifted from reduction of mortality towards longer-term complications of HIV infection and adverse effects associated with the use of antiretroviral therapy [²]. The adverse effects of antiretroviral drugs, including dyslipidemia, lipodystrophy, and insulin resistance are problems for the long-term management of HIV infection. This is of particular importance in children because HIV infection has become a chronic disease in this population. HIV-infected children may live for two decades longer than HIV-infected adults. As in the adult population, metabolic toxicity of antiretroviral therapy has been observed in children [^3-5], but the long-term cardiovascular consequences in children are unknown. Atherogenic lipoprotein changes in adults treated with protease inhibitors (PI) have been found and are associated with endothelial dysfunction and increased intima-media thickness (IMT) [^6-9]. However, the relative contributions of antiretroviral therapy, chronic inflammation due to the viral infection, classic cardiovascular risk factors, and their interactions are very difficult to identify. Bozzette et al. showed that the benefit in terms of mortality associated with the extensive use of therapies for HIV was not diminished by any increase in the rate of cardiovascular or cerebrovascular events or related mortality [¹⁰]. The paediatric population presents the unique opportunity to study vascular function during HIV infection in the absence of classic cardiovascular risk factors. Symptomatic atherosclerosis is evidently absent at this age, but endothelial function and arterial stiffness can be investigated by non-invasive echotracking techniques. As impairment of vascular function predicts future adverse cardiovascular events, we sought to determine if HIV-infected children receiving or not receiving antiretroviral therapy had early vascular dysfunction in comparison to healthy controls.

Methods

Patients

Forty-nine HIV-infected children (age range, 3.5-19.5 years) were recruited for the study from the cohort of HIV-infected children of our Institution and compared to 24 healthy controls matched for age and sex selected from our outpatient Department of Paediatric Cardiology among children evaluated for anorganic cardiac murmur. All were normotensive, had a normal electrocardiogram and echocardiography. None of these control subjects had known vascular risk factors. Five of the patients had neonatal blood transfusion-related infection and 44 acquired HIV by mother-to-child transmission. Thirty-two children received highly active antiretroviral therapy (at least three antiretroviral drugs of at least two different classes). Two children received only a combination of two nucleoside analogues. Fifteen had never been treated. All but one child were asymptomatic with more than 20% CD4 lymphocytes at the time of the study. The clinical and biological conditions, as defined in the PENTA guidelines for the use of antiretroviral therapy in paediatric HIV infection [¹¹], as justifying antiretroviral therapy being deferred were fulfilled in the cases of the never-treated patients. Subject characteristics are summarized in (Table 1). The median duration of treatment in 34 treated children was 96 months (range, 29-177 months) and median duration of PI treatment in 27 PI-treated patients was 45 months (range, 6-74 months). Our institutional review board approved the study. Informed written consent was obtained from the older children and from parents of all children. All the investigations were performed between November 2001 and May 2003 by two investigators (D.B. and Y.A.).

Table 1

Concentrations of total cholesterol and triglycerides in all HIV-infected patients were measured by enzymatic assay after an overnight fast. Dyslipidemia was considered present if total cholesterol and/or triglycerides levels were above the 95^th percentile for age and sex according to Christensen et al. [¹²].

Arterial measurements

Subjects were examined at a controlled room temperature of 22 ± 1°C. Blood pressure and heart rate were measured on the left arm, after the subjects had been lying supine for 10 min, automatically by an oscillometric recorder (Dynamap model 8100, Critikon).

All non-invasive measurements were performed by using a real-time B-mode ultrasound imager (Acuson XP 128, Mountain View, California, USA). IMT and lumen diastolic and systolic diameters were measured at the common carotid artery and mechanical parameters were calculated according to a previously described procedure [¹³]. After at least 30 min of rest, flow-mediated dilation defined as changes in arterial diameter in response to reactive hyperaemia (increased flow producing endothelium-dependent vasodilation) and glyceryltrinitrate-mediated dilation defined as changes in arterial diameter in response to the endothelium-independent vasodilator glyceryltrinitrate were studied with a high-resolution vascular ultrasound system as described previously [¹⁴]. The reactivity of the right brachial artery was satisfactorily assessed in 46 HIV-infected children and 24 controls. Repeatability of measurements was assessed as previously described [¹⁴]. The left ventricular mass index was calculated as previously described [¹⁵].

Statistics

Descriptive data are expressed as means and standard deviations (SD) or standard errors (SE) when appropriate. The characteristics of HIV-infected patients and controls were compared using an unpaired t test for continuous variables. Variables that were not normally distributed were log-transformed. Differences in categorical variables were evaluated by χ² tests. Univariate regression was used to assess correlations between variables. P < 0.05 was considered significant for all tests.

Results

Vascular function

The IMT in HIV-infected children was not different from that in controls. Cross-sectional compliance and distensibility were significantly lower in HIV-infected children. Diastolic wall stress and incremental elastic modulus were significantly higher in HIV-infected children, indicating an increased arterial stiffness (Table 1). There was no difference in resting diameters of the brachial artery in HIV-infected children and in controls (3.0 ± 0.4 versus 2.9 ± 0.5 mm, respectively). The flow mediated (endothelium-dependent) dilation was significantly lower in the HIV-infected children than in controls but the glyceryltrinitrate-mediated (endothelium-independent) dilation did not differ between the two groups, indicating endothelial dysfunction (Table 1).

The left ventricular mass index was slightly but not significantly higher in HIV-infected children than in the control group (111 ± 32 versus. 89 ± 28 g/m², P = 0.2).

There was no difference for any of the arterial variables between HIV-infected children receiving antiretroviral therapy and never-treated HIV-infected children (Table 2). Dyslipidemia was present in 17/49 HIV-infected children. Fifteen of the 27 patients receiving PI had elevated cholesterol and/or triglyceride levels and only two patients treated without PI had dyslipidemia (P < 0.001). Lipid values were consistently below the 95^th percentile in never-treated patients. All arterial variables were similar in patients with and without dyslipidemia (data not shown). There was no correlation between arterial variables and lipid values in patients with dyslipidemia.

Table 2

Discussion

Here, we show that HIV-infected children have a vascular dysfunction that may be an early step in the development of atherosclerosis. We did not find any difference between children receiving antiretroviral therapy and never-treated patients. This was surprising but the cross-sectional design of our study and the size of our population make it difficult to draw firm conclusions.

The effects of antiretroviral therapy on dyslipidemia and atherosclerosis in adult patients is currently an area of active investigation [^2,6-10]. To date, clinical studies have been unable to distinguish clearly between the contributions of 'classic' cardiovascular risk factors and of antiretroviral therapy in promoting atherosclerosis. Consequently, the relationship between antiretroviral therapy and premature atherosclerosis remains unclear [^7,8,10].

In our study none of the vascular variables differed between treated and never-treated children. Differences with the controls indicate that the HIV infection itself may have a deleterious effect on vascular function. This is consistent with autopsy studies during the pre-antiretroviral era reporting eccentric atherosclerosis lesions in the absence of traditional risk factors [¹⁶]. Joshi et al. reported a distinct arteriopathy involving small and medium sized arteries in paediatric AIDS patients [¹⁷]. The HIV envelope protein, gp120, activates human arterial smooth muscle cells to express tissue factor, the initiator of the coagulation cascade [¹⁸]; also inflammatory cytokines and viral proteins synergistically promote endothelial activation, apoptosis or cell proliferation [¹⁹]. Consequently, the arterial remodelling observed in never-treated patients could be a result of direct viral infection, or of the activation of bystander cells (smooth muscle cells and endothelial cells) by HIV viral proteins.

We showed that the carotid arterial wall was stiffer in HIV-infected children than in control subjects of the same age. Structural changes of the arterial wall occur very early in atherosclerosis. Although the IMT in our patients was normal, diastolic wall stress was significantly increased as was the incremental elastic modulus. These findings suggest a rearrangement of the arterial wall in HIV-infected children. Possibly, arterial wall remodelling is incipient in HIV-infected children, and is sufficiently advanced to increase vessel stiffness but not to increase the IMT. Alterations in endothelial function and, specifically, in flow-mediated dilation as in our patients, could account for altered vasomotor tone and/or arterial wall structure and therefore for properties of the conduit arteries. Arterial stiffness affects the ability of the large arteries to act as a cushion for cardiac output and is an important determinant of the vascular load to the heart. For this reason, increased arterial stiffness could explain, at least in part, the increased left ventricular mass index described by Lipshultz et al. in HIV-infected children [²⁰]. In our study, left ventricular mass index was moderately elevated in HIV-infected children but the difference from controls was not statistically significant. Possibly, antiretroviral therapy had a favourable effect on cardiovascular status in HIV-infected children as compared to the population of the Pediatric Pulmonary and Cardiovascular Complications of Vertically Transmitted HIV infection (P²C²HIVstudy) that was recruited between May 1990 and January 1994. It is thus possible that antiretroviral therapy counterbalances, at least transiently, HIV-induced injury to the developing vascular bed in children by reversing or stabilizing the HIV-induced vascular dysfunction. This would be consistent with the absence of a difference between treated and never-treated patients. The cross-sectional nature of our study does not allow us to confirm this possibility or assess the time-course of any vascular lesions associated with either HIV-infection or antiretroviral therapy. Longitudinal studies are needed to clarify these issues. Interestingly, Wolf et al. showed that antiretroviral therapy reduced markers of endothelial and coagulation activation in HIV-infected patients [²¹]. This beneficial effect might perhaps be transient if children are chronically exposed to adverse metabolic effects of antiretroviral therapy. It would be also of a great value to study vascular function in patients who were treated with antiretroviral therapy during the first months of life, to reveal whether this strategy protects them from the detrimental effects of HIV infection on vascular function.

Acknowledgments

We thank M.C. Mourey and I. Descamps for excellent secretarial assistance.

Sponsorship: Supported by the Agence National de Recherches sur le SIDA, France.